Sage Journals: Discover world-class research

Abstract

The novel coronavirus disease (COVID-19) has spread over 219 countries of the globe as a pandemic, creating alarming impacts on health care, socioeconomic environments, and international relationships. The principal objective of the study is to provide the current technological aspects of artificial intelligence (AI) and other relevant technologies and their implications for confronting COVID-19 and preventing the pandemic’s dreadful effects. This article presents AI approaches that have significant contributions in the fields of health care, then highlights and categorizes their applications in confronting COVID-19, such as detection and diagnosis, data analysis and treatment procedures, research and drug development, social control and services, and the prediction of outbreaks. The study addresses the link between the technologies and the epidemics as well as the potential impacts of technology in health care with the introduction of machine learning and natural language processing tools. It is expected that this comprehensive study will support researchers in modeling health care systems and drive further studies in advanced technologies. Finally, we propose future directions in research and conclude that persuasive AI strategies, probabilistic models, and supervised learning are required to tackle future pandemic challenges.

Keywords

artificial intelligence (AI)coronavirus disease (COVID-19)deep learning (DL)health care machine learning (ML) technology

The year 2020 started with the advancement of numerous digital technologies beneficial for health care. These technologies¹—including the internet of things (IoT) with fifth generation (5G) networks,^2,3 bigdata,⁴ artificial intelligence (AI), including machine learning (ML) and deep learning (DL),^5,6 and blockchain technology⁷—are used to tackle problems in traditional health care systems and the pandemic.⁸ The world is now challenging a global health catastrophe instigated by the coronavirus disease (COVID-19).⁹ COVID-19, initiated by the novel coronavirus (severe acute respiratory syndrome coronavirus-2 [SARS-CoV-2]), was identified by the World Health Organization (WHO) in December 2019 in China.¹⁰ The WHO acknowledged the Chinese outbreak of COVID-19 as a public health emergency on January 30, 2020, and declared a global pandemic on March 11, 2020,¹⁰ that posed a serious hazard for public health systems. The COVID-19 pandemic is now moving like a wave and has affected 219 countries, with more than 2.8 million deaths in the world, reported on April 5, 2021.^11,12 The pandemic has globally created negative impacts on socioeconomic, health, and political environments. The ways of tackling the spread of the coronavirus are initial identification, isolation, rapid management, spread prediction, and execution of systems for contact tracing.¹³ However, the major challenges are delays of virus tests, drugs, or medicines and providing services to critical zones. The main goal now is early detection and diagnosis of the virus, continuous checking and nursing of contacts, analysis of epidemiological and medical reports from patients, and progress of treatment procedures and drugs. In the fight against the coronavirus, Industry 4.0 technologies including AI, 5G-based IoT devices, and other digital technologies are vital for health, social, and economic performance. These technologies possess the competence of providing enhanced digital solutions for tackling the challenges throughout the calamity^14–17 and mitigating the global health crisis of this epidemic.^1,7,18,19

AI is one of the prospective technologies in health care for better understanding and addressing the COVID-19 crisis.^20,21 For the present pandemic purposes, AI can be defined as ML or DL, natural language processing (NLP), and computer vision applications that catalyze to build effective systems using big data for the global health emergency. These applications include: (a) AI-based clinical data analysis and robotic and remote surgery,²² (b) research and drug discovery,²³ and (c) automatic health care activities, processes, and management.²⁴ This article addresses the present status of AI, as well as the potential of AI approaches (including ML, DL, and NLP) in health care. It provides a comprehensive study of these AI approaches for better understanding and describing the COVID-19 pandemic crisis.

Background

In the past few decades, human civilization has faced several worsening pandemics and global public health emergencies. Since 2009, the WHO has declared 6 public health emergencies of international concern owing to global spread of diseases of the 2009 H1N1 (Swine flu) epidemic,²⁵ the 2014 polio outbreak,²⁶ the 2014 Ebola virus disease in Western Africa,²⁷ the 2016 Zika virus epidemic,²⁸ the ongoing 2018 Kivu Ebola epidemic,²⁹ and the COVID-19 pandemic.³⁰ The 2019 novel coronavirus (2019-nCoV) outbreak is the newest among the major epidemics and, thus, the WHO declared a health emergency due to the biggest viral disease outbreaks. The eruption of COVID-19 triggering novel-coronavirus-infected-pneumonia has hampered the lives of 132 million people in the globe, reported on April 4, 2021, according to the WHO.^11,12,30 The 2019-nCoV is now rapidly grown and distressing 219 countries and territories and 2 international conveyances (Figure 1), according to the Worldometer, updated on April 4, 2021.¹¹ The WHO officially renamed the disease as COVID-19 on February 11, 2020, and the virus that causes the disease as SARS-CoV-2 and declared a global health emergency. Researchers suggested a significant genomic similarity of 2019-nCoV with the severe acute respiratory syndrome (SARS) virus that appeared as a pandemic in 2002.³¹

Figure 1.

Illustration of the geographical spread of confirmed coronavirus disease (COVID-19): the highlighted areas are affected.

In dealing with the COVID-19 global pandemic, AI and related technologies have potential to help human societies and health care. In health care, AI and ML tools provide significant opportunities in the clinical areas of diabetes, heart disease, cancer, neurological disease, and infectious disease (see Figure 2).³² These technologies can be introduced to design efficient systems and to support researchers and physicians in addressing and overcoming challenges of the infectious coronavirus crisis. Various companies have launched a series of AI technologies to support organizations in the fight against COVID-19, such as estimation of outbreak size, coronavirus detection, diagnosis, analysis data and treatment procedures, research and drug development, and the future prediction of outbreak. AI is not one technology, but rather a group of technologies,^33,34 which can make significant contributions in the field of health care (Figure 3).

Figure 2.

Major clinical areas where AI and ML hold the most promise.

Figure 3.

Various aspects of artificial intelligence (AI) technology in health care.

Machine Learning and Deep Learning

The most important form of AI is ML, which has uses for various contexts in health care, such as diagnosing patient attributes, predicting treatment procedures, and precision medicine.³⁵ The composite forms of ML include DL or neural network models used in clinical data analysis and detection of particular diseases.^36,37

Natural Language Processing

The applications of NLP include design, construction, identification, understanding, and classification of clinical documents as well as research papers. The NLP can scrutinize unstructured medical information about patients, organize reports from radiological investigations, transcribe patient interactions, and extract information from unstructured data to enrich structured medical data.^38,39 It is combined with the ML-based knowledge graph to find a connection between the coronavirus and potential drug candidates and to provide outputs.^40,41

Expert Systems

Expert systems, involving human experts and knowledge engineers, are widely employed for clinical decision making³⁷ and electronic medical records providers prepare a set of rules with daily procedures in health care.

Robots and Robotic Process

Robots are used to deliver medical supplies in hospitals and other quarantine areas. AI-capable surgical robots can provide superpower to surgeons for improving their abilities; regular surgical measures employing robotic surgery²² include gynecologic surgery, prostate surgery, and head and neck surgery, among others. Robotic process automation does not actually involve robots, rather utilizing computer programs to perform repetitive tasks, such as authorization, modernizing medical histories, or billing processes in health care.³³

There are several potential contributions of AI against the coronavirus pandemic, such as early detecting, tracing, and forecasting, diagnosis and projection, treatments and medications, and social management and services,²¹ which are categorized in the following subsections.

Understanding and Detecting

To prevent the spread of infectious diseases, viruses should be tracked earlier. Quick detection of virus helps for patient isolation, contact tracing, treatment, and the delivery of emergency alerts to others. AI-based systems can learn to detect virus outbreaks by analyzing news reports, social media posts, and other outbreak-related documents. AI system BlueDot,⁴² using ML algorithms, was used to track and detect infectious disease outbreaks in Wuhan, China. Another AI predictive system, ClosedLoop C-19 index,⁴³ using expert knowledge in health care can identify those people who are at the highest risk of critical complications from COVID-19. AI-based surveillance systems, using facial recognition tools and temperature sensors, can identify people who have fevers and who have coronavirus. Several search engines, social networking sites, and media sites using ML approaches can be used to track flu-related disease outbreaks in real time.⁴⁴ But the main challenge here is data safety and security. It is required to establish a standard protocol to permit communication between devices and systems without compromising data integrity.

Diagnosis, Genome Analysis, and Treatment

Quick and correct findings of COVID-19 cases can save lives and prevent further outbreaks. AI technology has a superior impact on faster disease identification and diagnosis.²⁰ It can extract information from various clinical reports, analyze DNA sequences, and advise on treatment procedures.⁴⁵ A DL method⁴⁶ and a supervised neural network model (known as COVNet)^47,48 were developed to examine the chest x-ray and computed tomography (CT) scan images of patients and identify COVID-19 cases.⁴⁹ The model can accurately distinguish between COVID-19(+) and COVID-19(−) patients.⁵⁰ Multiple AI companies have launched AI-powered diagnosis systems that can help physicians diagnose the virus and monitor disease outbreaks. In the United States, XR Health⁵¹ recently announced a virtual reality telehealth support group to assist persons in isolation due to the virus, so that patients with similar diseases can get support from each other and from physicians. Also, DL approaches are applied in SARS-CoV-2 genomic sequencing that can process large and complex genomic datasets.^52–54 For instance, Google DeepMind's AI algorithms can help to understand the protein structure of the virus and support the development of treatment procedures for new diseases.⁵⁵ The AI systems for diagnosis, analysis, and treatment face logistical difficulties in implementation, because of dataset swing, new populations, and the negative consequences of new algorithms on health outcomes.

Prediction and Forecasting the Outbreak

AI has been employed for keeping track of and predicting how COVID-19 spreads over time and space.²¹ Akhtar et al⁵⁶ presented a dynamic artificial neural network prototype to forecast the span of the COVID-19 pandemic. This approach was applied for the prediction of the 2015 Zika virus pandemic. HealthMap⁵⁷ and BlueDot⁴² have been developed using ML algorithms that can efficiently predict the outbreak of the virus. Wang et al⁵⁸ proposed a prognosis model for influenza diagnosis, depending on analysis of the real-time Twitter information that can help to prevent future outbreaks.⁵⁹ The XGBoost model⁶⁰ is another ML-based prediction model that was used to diagnose a patient disease caused by the coronavirus. The model is convenient for analyzing whether a person should be isolated due to the COVID-19. Several studies have been performed to gather training data from the present pandemic and to develop a precise prediction model for the pandemic.⁶¹ Google Flu Trends⁶² presented an approach that employs online searching for estimating and predicting real-time flu activity using AI and big social data. This forecasting and prediction enable health officials to make better response plans and manage the pandemic.⁶³ Lacking sufficient datasets, AI-based forecasting models may not represent true transmission patterns.

Speeding up Research and Drug Discovery

Amazon Web Services (AWS) developed a new search engine, CORD-19,⁶⁴ using ML and NLP that can help scientists and researchers search huge volumes of research articles and documents quickly. In California, an AI-based company, C3.ai,⁶⁵ recently founded a research consortium aimed at tackling COVID-19 by utilizing AI and related strategies in finding the virus spread, forecasting its evolution, developing new drugs, and fighting future outbreaks.⁶⁶ AI and ML can identify and predict viral proteins of 2019-nCoV for drug development, support that can speed up the discovery of medications and vaccines to treat COVID-19.^54,55,67 An AI company in the United Kingdom, BenevolentAI,⁶⁸ aimed at discovering drugs and immediate medicines for COVID-19 by employing AI, DL, and other cutting-edge technologies. Even before the COVID-19 outbreak, AI had potential contributions to health care for new drug discovery.^69,70 Generally, AI is still in its initial stages in this area, and the prediction accuracy of ML methodologies, together with more real-world relevance, will continue to develop.

Public Health Management and Services

Several AI-based activities have the potential to contribute to management of the pandemic, such as helping doctors to monitor health crises and handle multiple patients in hospitals, scanning infected people in public spaces using thermal imaging, measuring social distancing and lockdown procedures, and providing health consultation services to raise awareness around the world. Also, AI-based systems and virtual reality technology provide social services for COVID-19 patients and help to educate others. Such systems, therefore, illustrate how society and public administrations can benefit from the use of AI technology. AI-based drones are used for safe and fast delivery of medical supplies in critical zones⁷¹; AI-enabled robots are used for cleaning, sterilizing, and delivering food and medicine to avoid human-to-human contact; and ML-based UVD robots using ultraviolet light can disinfect and autonomously kill bacteria and viruses to limit the spread of coronavirus. The major threats against the safety of the people and AI devices as well as their interference with air traffic are needed to be addressed.

Related Datasets

Data is a vital factor in helping public, governmental, and health care sectors combat the COVID-19 pandemic. It is key to remember that not all data is complete and accurate. Thus, it is necessary to handle uncertainty and inconsistency in coronavirus datasets. Very recently, different datasets have been used by researchers and scientists for their research related to COVID-19. A few of the used datasets related to the coronavirus pandemic are listed below.

The GISAID data⁷² from all influenza viruses and the coronavirus, including virus genomic sequence and related medical and epidemiological data, used for predicting protein structures and RNA sequencing.⁷³

Coronavirus genome sequence dataset in the Kaggle repository⁷⁴ used for RNA sequencing of a taster of bronchoalveolar lavage fluid from a patient identified by a new RNA virus injure from the family of Coronaviridae named as 2019-nCoV.⁷³

CORD-19 dataset comprising more than 29 000 academic papers on COVID-19, SARS, Middle East respiratory syndrome (MERS), and allied viruses in the Kaggle repository,⁷⁵ used to extract the most relevant information about the disease.

GitHub coronavirus repository⁷⁶ provides the daily COVID-19 case count files, and all data operations are vectorized, from which users can generate new CSV, JSON, or Pickle files.

COVID-19 Korean dataset⁷⁷ with patient routes and patient age/gender/diagnosed date on the disease from Korea, used for data visualization with features of displaying infected patient routes and regional patient count.

COVID-19 image dataset,⁷⁸ a public dataset in GitHub repository, contains chest x-rays and CT scans of patients who are affected by COVID-19 or other viral and bacterial pneumonia, such as MERS, SARS, and acute respiratory distress syndrome. Data are gathered from different open sources, clinics, and physicians and used to design AI-based applications to understand and detect contamination.

Coronavirus Tweets dataset in Kaggle repository⁷⁹ contains the Twitter data of the hashtags: #coronavirus, #coronavirusoutbreak, #coronavirusPandemic, #covid19, #covid19, #epitwitter, and #ihavecorona, and is utilized to forecast coronavirus outbreaks.

COVID-19 coronavirus dataset in European Union Open Data Portal⁸⁰ is public data on the COVID-19 pandemic, including daily updates and worldwide cases. The dataset is published by the European Center for Disease Prevention and Control.

COVID-19 pandemic data in the Hum Data repository⁸¹ shows global COVID-19 data with confirmed cases and deaths in places with humanitarian response plans; data is sourced from the WHO.

The novel coronavirus (COVID-19) cases data in the Hum Data repository⁸² is COVID-19 epidemiological data compiled by the Center for Systems Science and Engineering of Johns Hopkins University⁸³ from various sources, including the WHO and others. The dataset contains fields of last update, confirmed, suspected, recovered, deaths, and locations.

COVID-19 dataset in Github repository⁸⁴ is a set of COVID-19 pandemic data, updated daily and including up-to-date data on confirmed cases, deaths, and testing, throughout the duration of the pandemic.

Coronavirus (COVID-19) Tweets dataset in IEEE Data Port⁸⁵ contains real-time Twitter data for coronavirus-related tweets using keywords of corona, coronavirus, COVID, pandemic, lockdown, quarantine, social distancing, N95, SARS-CoV-2, 2019-nCov, COVID-19, etc.

Coronavirus disease (COVID-19) dashboard²⁹ powered by the WHO provides up-to-date pandemic data (new cases, confirmed cases, and deaths) from around the world and visually displays the latest global coronavirus outbreak on the world map.

COVID-19 coronavirus pandemic analyzer, powered by Worldometer,⁸⁶ analyzes and aggregates worldwide pandemic data from thousands of sources in real time. This analyzer visually displays global outbreak graphs according to dates and location.

The COVID-19 Open Research Dataset used CORD-19,⁶⁴ which was developed by AWS. It receives natural language questions and provides the answers from Amazon Kendra, with the dataset updated regularly. Chest x-ray images (pneumonia) dataset⁸⁷ consists of 5863 JPEG images with 2 classes, pneumonia and normal, available on the Kaggle repository. These images were collected from early aged patients with pneumonia from the Guangzhou Women and Children’s Medical Center, China.

ImageNet database⁸⁸ contains more than 14 million images across 20 000 categories that were generated for image analysis competitions⁸⁹ by Stanford Vision Lab at Stanford University and Princeton University.

CATNAP database⁹⁰ consists of antigen and antibody structures of viruses, such as human immunodeficiency virus (HIV), H1N1, Dengue, SARS, and Ebola. The CATNAP tools have been developed to promote the investigation of neutralizing antibodies over the detection of prospective genetic signatures.⁹¹

Protein data bank,⁹² funded by the U.S. National Science Foundation and the National Institutes of Health, provides data about the 3-dimensional patterns of proteins and nucleic acids and helps researchers synthesize protein patterns for understanding the diseases.

COVID-19 data lake⁹³ is a public, up-to-date database, where anyone can experiment with and analyze real-time coronavirus pandemic data worldwide (confirmed cases and deaths), according to country and date.

Artificial Intelligence Approaches in Health Care

AI methods can be categorized into 2 main classes in health care applications: ML, including DL, and NLP approaches. The ML and DL strategies^94,95 investigate structured data, such as imaging, genetic, and electronic medical records data, and attempt to classify patients’ characteristics or deduce the possibility of infection consequences.³⁹ NLP⁹⁶ procedures collect data from unstructured sources, such as medical notes, reports, or journal data, to enrich the structured data.³⁸ This study summarizes different AI models that can be used to detect and diagnose coronavirus infections, monitor the spread of COVID-19 outbreaks, and provide various health care services.

Machine Learning Models for Detection and Diagnosis

ML architectures, such as convolutional neural networks (CNNs), provide widespread applications in medical image analysis for feature selection, classification, segmentation, and lesion detection for medical data.^97–100 Medical imaging and signal data, such as x-ray, CT, and magnetic resonance imaging, are analyzed with the help of DL models. These analysis results have a great impact on the detection, identification, and diagnosis of several diseases, such as cancer, brain tumor, diabetes, and pneumonia.^50,101–105 Recently, the reverse transcription–polymerase chain reaction is a standard test method used in detecting and categorizing respiratory viruses.¹⁰⁶ Although this standard method of diagnosis is capable of detecting COVID-19, computer-aided ML techniques nevertheless aid faster detection and diagnosis. Accordingly, rapid and cost-effective solutions for fast COVID-19 detection utilize AI- and ML-based techniques^107,108 for clinical image processing on the coronavirus.^109–113 In most of the research, x-ray and CT scan images are used as the input of ML and DL prototypes for early detection of virus-infected cases.^{49,50,114,115}

Different pretrained DL models, such as ResNet50, InceptionV3, VGG-19, and Inception ResNetV2, have been presented for detecting COVID-19-infected patients by employing chest x-ray and CT scan images.^106,112,115 For the experiment, COVID-19 patients’ imaging data have been collected from the GitHub⁷⁸ and the Kaggle⁸⁷ public repositories. The residual neural network (ResNet) architecture is an improved form of the CNN model¹¹⁶ and ResNet50 is a 50-layer neural network-based model employing a scheme of skip links between layers, known as residual learning and trained on the ImageNet dataset.⁸⁹ InceptionV3,¹¹⁷ one kind of CNN model that contains a fully connected neural network, improves the use of manipulated resources inside the network. The Xception¹¹⁸ CNN model developed by Google was an updated form of the Inception model. VGG-19,¹¹⁹ having 19 layers of deep neural network architectures with low convolution filters, was modeled to reach high accuracy in health care applications with a high volume of image data.¹²⁰ Another pretrained architecture contains a deep neural network employing the Inception ResNetV2 model that was trained on the patients’ image dataset and provided the output as a list of estimated class probabilities of the diseases.¹²¹ The schematic illustration of conventional CNN, including pretrained ResNet50, InceptionV3, VGG-19, and Inception ResNetV2 architectures for COVID-19 detection and diagnosis, is shown in Figure 4. It provides the 3 types of detection, such as COVID-19(+), COVID-19(−), and others. This model helps health staff determine which cases should be investigated with standard methods and which management approaches should be employed in distinguishing COVID-19 and non-COVID-19 cases.

Figure 4.

DL-based frameworks for COVID-19 detection and diagnosis.

Narin et al¹¹² presented 5 pretrained DL models, such as ResNet50, ResNet101, ResNet152, InceptionV3, and Inception-ResNetV2, for finding COVID-19 patients using chest x-ray images. The datasets contain the chest x-ray images of bacterial and viral pneumonia, comprising coronavirus and other cases as well as data from healthy people, using data acquired from the GitHub⁷⁸ and Kaggle⁸⁷ open repositories. They investigated the detection success of COVID-19 (+), COVID-19 (−), or other cases, and the performances of these 5 models on both training and testing datasets are illustrated in Figures 5 and 6. The ResNet50 model provided the supreme average success rate of 98.43% for all datasets, whereas the InceptionV3 model had a detection performance of 100% in only the COVID-19 dataset. Haque et al¹²² proposed a CNN-based model that identified only COVID-19 patients using the chest x-ray images, and this approach achieved an accuracy rate of 97.56%. The pretrained ResNet50 proposed by Song et al¹²³ achieved a success rate of 86% using CT images. Another study proposed by Xu et al¹²⁴ achieved an accuracy of 86.7% in identifying COVID-19 cases utilizing the ResNet model. Wang et al⁴⁹ used CT images in the modified inception model and achieved a sorting precision of 82.9%. In the study,¹²⁵ a deep neural network model was implemented that offered the precise diagnostics for multiclustering, such as COVID-19(+), pneumonia, and none, using chest x-ray images. This approach applied the combined ResNet50–support vector machine (SVM) models and performed the classification with 95.38% accuracy. Although ML-based imaging approaches show a significant role in the analysis of COVID-19 cases, the irregularities in datasets remain the prime challenge in COVID-19 diagnosis from image data.

Figure 5.

The comparison plots of training accuracy of the 5 pretrained convolutional neural network (CNN)-based models.¹¹²

Figure 6.

The comparison plots of testing accuracy of the 5 pretrained convolutional neural network (CNN)-based models.¹¹²

Machine Learning and Natural Language Processing Model for Clinical Decision Making and Management

Data collected from various sources are not always structured and require NLP procedures to extract data features from disorganized data sources to enrich structured medical data.³⁸ ML-based approaches are then applied to analyze structured clinical data, provide the clustering of patients’ reports, and subsequently assess the probability of disease outcomes.³⁹ Thus, NLP procedures generate machine-readable structured and clustered data from texts, and then the clustered data can be analyzed by ML techniques to provide outputs, as shown in Figure 7.¹²⁶ The figure illustrates the process flow of structured data generation from clinical documents, including images, genetic data, electronic medical records, and electrophysiological data, through NLP and ML, to suggest clinical activities.

Figure 7.

AI-based clinical models, including NLP-based data enrichment and ML-based data analysis.¹²²

AI-based models can be applied to solve clinical problems and to assist medical practices in primary detection, identification, diagnosis, treatment, and management of critical COVID-19 cases.^127–129 Another AI-based model is shown in Figure 8. DL (or deep neural network) approaches could be useful tools for the differential diagnosis of COVID-19, and the genomic variants from normal or severe COVID-19 patients can be classified by the ML analyzer to predict potential COVID-19-infected patients.⁵³ This model could also be used to monitor critically ill patients and assist physicians in decision making about further therapy. Although ML models provide significant improvements in clinical decision making, a large volume of studies is needed to ensure these outcomes and to increase their acceptance in health care sectors.

Figure 8.

Artificial intelligence (AI) models for clinical decision making and management.

Artificial Intelligence System for Identifying COVID-19 Symptoms

AI technology can be conveniently employed for keeping track of the spread of virus outbreaks, identifying extremely critical patients and regions, and controlling virus infections expeditiously. AI can forecast the death risk and recovery cases by analyzing patients’ earlier data. Figure 9 illustrates the methodological steps of an AI-based system that can help medical staff to identify the COVID-19 symptoms and suggest the necessary treatments for COVID-19 patients.¹³⁰ The system starts with COVID-19 symptoms and patients’ test analysis. After confirming COVID-19-infected cases with the aid of AI, doctors decide further therapy and start treatment procedures and monitoring. The doctors are paying attention to treatment of the patients as well as AI applications to administer the spread of the infection. Thus, AI is a useful tool to identify early infections of COVID-19 and to improve treatment consistency by continuously monitoring infected patients using ML algorithms.

Figure 9.

General procedure of AI-based approach to identify COVID-19 symptoms.

Machine Learning-based Forecasting Methods

Fong et al⁶³ presented an ML methodology with a small dataset and conducted an investigation with the recent coronavirus eruption found in Hubei, China. It was an optimized forecasting model with a polynomial neural network that was able to forecast with a relatively low prediction error. The prediction architecture was designed with 5 classes of data analytics, and these classes embraced different types of ML algorithms.¹³¹

Class 01: Prediction employing ML architectures with manifold regression analysis. These models include some extent of neural network architectures that possess extremely configurable organizations and parameters, applied to multivariate datasets.¹³² With the perspectives of initial stages of epidemic prediction where long dimensional data are not obtainable, these models would be required for reducing forecasting errors.

Class 02: Prediction employing complex ML architectures. These models employ comparatively complicated ML processes, such as SVM or neural networks, which involve iterative hyperplanes or weights modifications for minimizing errors. Hence, these complex models consume significant time and space complexity.

Class 03: Prediction employing lightweight ML architectures. They are more reliable ML models, which run faster and produce results as conventional time series predictions.

Class 04: Simple data analytics. Most are traditional statistics, such as mean, median, minimum, maximum, increment, rate of increases of suspected cases, rate of recovery, etc.

Class 05: Econometrics kind of time series prediction. Classical econometrics, such as autoregression, exponential regression, moving average, autoregressive integrated moving average and its variants, etc. These approaches typically are employed on univariate time series with nil or small parameters.

From the study, it is noted that traditional prediction from Class 05 can be initially used, possibly simultaneously with Class 04 statistical analysis. With available resources and enough time, the strategies from Classes 03, 02, and 01 can be used in forecasting models that provide relatively low prediction errors. Moreover, forecasting is a dynamic process whose estimated outcome is sensitive to the model parameters and the availability of data. Thus, it is crucial to test popular ML algorithms with complete datasets in the new field.

Machine Learning for Accelerating Drug Development

Because there are still no explicit medications available for COVID-19, it is crucial to determine quick and effective solutions to inhibit COVID-19 infection and virus incidence. Magar et al¹³³ introduced an AI-based model that combines big data and medical knowledge with ML to determine the antibody sequences that can prevent the progress of coronavirus.¹⁰⁶ A pictorial diagram of the proposed model is shown in Figure 10,¹³³ which consists of 4 major modules: training dataset, extracting features, learning, and selecting hypothetical candidates. This ML model was trained on the dataset consisting of 1831 antigen and antibody sequences of numerous viruses (such as H1N1, Dengue, SARS, Ebola, and HIV) from the CATNAP tool⁹¹ and 102 samples of the RCSP protein data bank⁹²; hence, the final dataset consists of 1933 samples.

Figure 10.

A data-driven framework for discovering antibody sequences to treat coronavirus disease (COVID-19).

An ML-based cheminformatics software, RDKit,¹³⁴ was used to represent the molecular graph and extract the features from the dataset of 1933 samples. Five ML frameworks, including random forest, XGBoost, multilayer perceptron, SVMs, and logistic regression,¹³³ have been applied to assess their performance, and the experimental results showed that the XGBoost architecture achieved the best performance to determine the impending antibody candidates for COVID-19. Because the COVID-19 virus has similar characteristics as SARS-CoV-2,¹³⁵ the researchers recommended producing 2589 hypothetical candidates based on neutralizing antibodies and sequence features against SARS, and among these candidates, eventually, 8 structures were taken as prospective antibody sequences for neutralizing COVID-19.

As COVID-19 is caused by a novel coronavirus, proper understanding of viral structures helps to accelerate drug discovery. The vaccine development for a new disease is expensive and time-consuming; ML approaches can help to accelerate the entire process efficiently.¹³⁶ As discussed earlier, ML with NLP methods allows for a rapid processing of extensive medical data and extracts the key information from the data that helps to enable new structural predictions and the repurposing of existing drugs. The study and results would be useful for finding effective medicines and vaccines for COVID-19. AI helps to speed up the entire pharmaceutical process, from disease prediction to drug development (Figure 11).⁵⁵ ML and DL methods have the potential to assist drug development and biological research, but some limitations should be considered, such as knowledge about the selection of appropriate models, need for large training datasets, and high computational costs of training.

Figure 11.

AI and ML methodologies in emerging disease.

Machine Learning-based Search Engine

An ML-based search engine launched by AWS, named CORD-19,⁶⁴ can help researchers find coronavirus-relevant documents using natural language questions. The CORD-19 search offers an easy search interface in which researchers can make queries using natural language and produces detailed answers as well as source documentation. Firstly, the NLP tool is used to enrich the dataset, and the ML method is used to mine pertinent health information from unstructured text data about disease, treatment, and related information. The relevant information is then mapped to COVID-19 medical models using inference procedure and multilevel classification, involving different enrichment sources, and indexed in Amazon Kendra,¹³⁷ which is a precise search provision based on ML and DL approaches. Figure 12 illustrates the overall architecture of the CORD-19 search model.¹³⁸ It has robust natural language query capabilities that make it convenient to find and rank associated documents. The Amazon search model is developed and built from public data, available on the AWS COVID-19 dataset.⁹³ In the vast information world, NLP-based intelligent search engines provide an appropriate platform in which users must only input text without having to make judgments for the search results. Thus, it produces the correct documents and saves time and effort for users.

Figure 12.

CORD-19 search architecture.

Discussion and Future Directions

A huge number of data has to be aggregated and analyzed to get useful information about the coronavirus. This aggregated data would be able to show the latest details about the onset and spread of the virus, as well as the latest findings related to the transmission of the disease. Thus, ML with big data analysis is used in health care to build a strong prediction tool for coronavirus outbreaks. On the other hand, NLP is used to extract information from unstructured data to enrich structured medical data. It is the matter of training a proper neural network on a prepared dataset that supports the machine in diagnosing the disease. ML models can also analyze chest x-ray and CT scans using deep neural networks that provide significant information on the patient's status, and the blood sample visual analysis can provide additional information on a patient's health and overall condition. Some companies have made efficient tools based on ML and DL to help health care professionals in detection and diagnosis of the virus, surveillance of contagious illnesses, and prediction of the virus outbreak efficiently, as well as researching and collecting a huge amount of data from various sources. Nowadays, the ML model with big data can be mined for the relevant information from social media for tracking the spread of coronavirus. Technology-based health care has great potential to improve patient outcomes, predict outbreaks of epidemics, forecast current virus infection, gain valuable insights, screen patients, map and prevent diseases, determine the chances of disease incidence, detect risk areas and stop the spread of the disease, deliver medical supplies, and foster the quality of daily life. According to the study, technology can help to accelerate research activities in health care as well as manage clinical and social services.

Two key points have to be noted in the literature, the datasets and the methodologies. The most crucial issue is that researchers have studied a limited volume of COVID-19 data in their studies. Increasing the data with many clusters will empower stable systems. The data were collected from many different sources, not a centralized repository. Both male and female cases were analyzed in COVID-19 datasets, in which a remarkable number of aged patients were available. No feature extraction and feature selection procedures were employed in many studies, and they utilized the methods directly with raw data. Moreover, the ML and DL approaches have potential in combating the COVID-19 pandemic with limited resources. Many emerging technology-based solutions can be suggested to improve the public safety, security, and privacy concerns during this pandemic and in the future. The advanced technologies, including blockchain,⁷ federated ML,¹³⁹ incentive mechanisms with big data,¹⁴⁰ combined AI and cloud computing,¹⁴¹ 5G wireless technologies,¹⁴² and others, can offer effective solutions for confronting the COVID-19 pandemic. Researchers attempted to develop an effective approach by combining various models and methodologies with various technologies. Although these models show encouraging outcomes, they need further improvements with large datasets and clinical trials for employment in the health care sector.

The following research directions could be taken with AI technology to further the fight against future pandemics such as COVID-19:

Developing specialized, AI-based IoT architectures with big data to solve epidemic-related issues and build a comprehensive system to respond to COVID-19-like crises, such as prognostic modeling, real-time patient observing, and emergency operations in health care.

Using emerging ML to support decision making, drug discovery and treatment, medical devices, and the overall modernization of the medical industry. Hospital management and health care professionals are considering the advantages of applying AI in technology and keeping patients’ information in private clouds, such as Google Cloud Platform. AI provides hospital management and doctors with ways to store patient data on clouds more securely and access health records via AI-infused technology.

Improving AI approaches for medical imaging analysis will help to examine patients’ radiological and pathological reports accurately by machine, with doctors making proper decisions for treatment. Speech and text recognition using NLP could be employed for patient–doctor communication and capturing of clinical notes in remote treatment.

Collecting centralized, worldwide COVID-19 datasets could be beneficial for technological research to develop smart systems for prediction, diagnosis, forecasting, transmissibility, pathogenicity, and therapeutic methods during pandemics.

Conclusion

The outbreak of COVID-19 is observed keenly and this study concerns the reasons, impacts, and prevention policies at the very early stage of the pandemic. The outcomes of the study are summarized and a profound alignment with technology, especially AI and ML, is explored. This investigation intends to provide the current technological aspects of ML strategies that are appropriate to use in health care for facing the current pandemic situation. This article addresses the link between the technologies and the epidemics as well as the potential impacts of technology in medical services with the introduction of ML, DL, and NLP tools. AI has played an important role in confronting the coronavirus pandemic and supported researchers in developing systems to reduce human interaction in infected places, deliver services, and manage health emergency situations, but it needs ample resourcing and financial support to respond to public health threats. AI-related technology is being applied to track the span of infection rapidly, and the developed systems have potential to predict future outbreaks of the virus by early screening or diagnosis of patient data as well as inform the legal and ethical challenges of drug development in response to public health emergencies. Privacy and security are important concerns in health care data, and it is imperative to protect individuals’ personal data through proper policies and authorization requirements. The designated methods should address the protection of data from pernicious attacks and ensuring that data is shared appropriately. The best, persuasive AI strategies, AI-based searching strategies, probabilistic models, and supervised learning are required to tackle future pandemic challenges and public health issues.

Footnotes

Acknowledgments

The authors wish to thank the Department of Computer Science and Engineering of the Jatiya Kabi Kazi Nazrul Islam University (@JKKNIU, Bangladesh) and the ISE of the Universiti Tenaga Nasional (@UNITEN, Malaysia) for their supports and cooperation in conducting the research.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Md Mijanur Rahman

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Md Mijanur Rahman is a professor of the Department of Computer Science and Engineering, Jatiya Kabi Kazi Nazrul Islam University, Trishal, Mymensingh, Bangladesh. Dr. Rahman obtained his BSc (2001) and MSc (2002) degrees, both with first class, in computer science and engineering from Islamic University, Kushtia, Bangladesh. His PhD degree was awarded in computer science and engineering from Jahangirnagar University, Savar, Dhaka, Bangladesh, in 2014. His teaching and research interests are in areas of AI and ML, digital speech and image processing, the internet of things, and biometrics.

Fatema Khatun is an assistant professor in the Department of Electronics and Telecommunication Engineering, Bangabandhu Sheikh Mujibur Rahman Science & Technology University, Gopalganj, Bangladesh. Now she is a PhD fellow in the Institute of Sustainable Energy of the Universiti Tenaga Nasional, Malaysia. Ms. Khatun obtained her BSc (2003) and MSc (2004) degrees, both with first class, in electrical and electronic engineering from Islamic University, Kushtia, Bangladesh. She has more than 10 years of academic experience in electrical and electronic engineering-related disciplines at various universities in Bangladesh. Her areas of teaching and research interests include AI technology, renewable energy, material science, and nanotechnology.

Ashik Uzzaman received his BSc degree in computer science and engineering from Jatiya Kabi Kazi Nazrul Islam University, Mymensingh, Bangladesh. He is currently working with Robotic Process Automation at Remote Office, Dhaka, Bangladesh. His academic and research interests include robotic technology, DL, data science, NLP, and computer vision.

Sadia Islam Sami received her BSc in computer science and engineering from Jatiya Kabi Kazi Nazrul Islam University, Mymensingh, Bangladesh. Her research interests include AI, internet of things (IoT), image processing, and computer vision.

Md Al-Amin Bhuiyan received his BS and MS degrees in applied physics and electronics from the University of Dhaka in 1987 and 1988, respectively. Dr. Bhuiyan received his PhD in electrical and electronic engineering from Osaka City University, Japan, in 2001. From 2001 to 2003, he was employed as a COE post-doctoral research fellow with the Intelligent Systems Research Laboratory, National Institute of Informatics, Japan. He did his second post-doctoral research at the University of Hull, UK. Since 2008, he has been a professor with the Computer Science and Engineering Department, Jahangirnagar University, Bangladesh. Currently, he is a faculty member in the Department of Computer Engineering, King Faisal University. His research interests include image processing, computer graphics, pattern recognition, computer vision, AI, neural networks and robotic vision.

Tiong Sieh Kiong is a professor in the College of Engineering and the director of the Institute of Sustainable Energy, Universiti Tenaga Nasional, Malaysia. He received his BEng (Hons), MSc, and PhD in electrical, electronic, and system engineering from the National University of Malaysia in 1997, 2000, and 2006, respectively. His research interests are in renewable energy, AI, data analytics, microcontroller system, and communication systems. He is a professional engineer registered with the Board of Engineers Malaysia. He is also a member of the Institute of Electrical and Electronic Engineers.

References

Ting

DSW

Carin

Dzau

, et al. Digital technology and COVID-19. Nat Med. 2020;26(4):459–461. https://doi.org/10.1038/s41591-020-0824-5

Perkel

. The internet of things comes to the lab. Nature. 2017;542(7639):125–126. https://doi.org/10.1038/542125a

Ting

Lin

Ruamviboonsuk

, et al. Artificial intelligence, the internet of things, and virtual clinics: ophthalmology at the digital translation forefront. Lancet Digit Health. 2019;2(1):E8–E9. https://doi.org/10.1016/S2589-7500(19)30217-1

Shilo

Rossman

Segal

. Axes of a revolution: challenges and promises of big data in healthcare. Nat Med. 2020;26(1):29–38. https://doi.org/10.1038/s41591-019-0727-5

LeCun

Bengio

Hinton

. Deep learning. Nature. 2015;521(7553):436–444. https://doi.org/10.1038/nature14539

Ting

DSW

Liu

Burlina

, et al. AI for medical imaging goes deep. Nat Med. 2018;24(5):539–540. https://doi.org/10.1038/s41591-018-0029-3

Nguyen

Dinh

Pathirana

Seneviratne

. Blockchain and AI-based solutions to combat coronavirus (COVID-19)-like epidemics: a survey. TechRxiv. Preprint. 2020:2020040325. https://doi.org/10.36227/techrxiv.12121962.v1

Editorial. Finding a role for AI in the pandemic. Nat Mach Intell. 2020;2:291. https://doi.org/10.1038/s42256-020-0196-z

Wang

Horby

Hayden

Gao

. A novel coronavirus outbreak of global health concern. Lancet. 2020;395(10223):470–473. https://doi.org/10.1016/S0140-6736(20)30185-9

10.

Sohrabi

Alsafi

O’Neill

, et al. World Health Organization declares global emergency: a review of the 2019 novel coronavirus (COVID-19). Int J Surg. 2020;76:71–76. https://doi.org/10.1016/j.ijsu.2020.02.034

11.

COVID-19 Coronavirus Pandemic. Accessed April 5, 2021. https://www.worldometers.info/coronavirus/

12.

HealthMap, COVID-19. Accessed April 05, 2021. https://www.healthmap.org/covid-19/

13.

World Health Organization. Novel Coronavirus (2019-nCoV): Situation Report 12; 2020. Accessed June 10, 2020. https://apps.who.int/iris/handle/10665/330777

14.

Cheng

Liu

Qiang

Liu

. Industry 4.0 development and application of intelligent manufacturing. In 2016 International Conference on Information System and Artificial Intelligence (ISAI), Hong Kong; 2016:407–410. https://doi.org/10.1109/ISAI.2016.0092

15.

Grasselli

Pesenti

Cecconi

. Critical care utilization for the COVID-19 outbreak in Lombardy, Italy: early experience and forecast during an emergency response. JAMA. 2020;323(16):1545–1546. https://doi.org/10.1001/jama.2020.4031

16.

Ahmed

Quadeer

McKay

. Preliminary identification of potential vaccine targets for the COVID-19 coronavirus (SARS-CoV-2) based on SARS-CoV immunological studies. Viruses. 2020;12(3):254. https://doi.org/10.3390/v12030254

17.

Javaid

Haleem

Vaishya

, et al. Industry 4.0 technologies and their applications in fighting COVID-19 pandemic. Diabetics Metab Syndr. 2020;14(4):419–422. https://doi.org/10.1016/j.dsx.2020.04.032

18.

Haleem

Javaid

. Additive manufacturing applications in Industry 4.0: a review. J Ind Integr Manage. 2019;4(4):1930001. https://doi.org/10.1142/S2424862219300011

19.

Ren

Zhang

Wang

. Traditional Chinese medicine for COVID-19 treatment. Pharmacol Res. 2020;155:104743. https://doi.org/10.1016/j.phrs.2020.104743

20.

Bullock

Luccioni

Pham

, et al. Mapping the landscape of artificial intelligence applications against COVID-19. March 25, 2020. https://arxiv.org/abs/2003.11336v2

21.

Naudé

. Artificial intelligence against COVID-19: an early review. IZA Discussion Paper 2020, No. 13110. https://covid-19.iza.org/publications/dp13110/

22.

Andras

Mazzone

van Leeuwen

FWB

, et al. Artificial intelligence and robotics: a combination that is changing the operating room. World J Urol. 2019;27:1–8. https://doi.org/10.1007/s00345-019-03037-6

23.

Senior

Jumper

Hassabis

Kohli

. Alphafold: using AI for scientific discovery. DeepMind News . January 15, 2020. Accessed July 12, 2020. https://deepmind.com/blog/article/AlphaFold-Using-AI-for-scientific-discovery

24.

Sarkar

. AI and machine learning for healthcare; a brief overview of AI/ML as being applied to healthcare. Medium Article . April 24, 2020. Accessed July 12, 2020. https://towardsdatascience.com/ai-and-machine-learning-for-healthcare-7a70fb3acb67

25.

Bautista

Chotpitayasunondh

Gao

, et al. Clinical aspects of pandemic 2009 influenza A (H1N1) virus infection. N Engl J Med. 2010;362(18):1708–1719. https://doi.org/10.1056/NEJMra1000449

26.

World Health Organization. Status Report on Progress Against the Polio Eradication and Endgame Strategic Plan 2013-2018. WHO publications; 2014. Accessed June 10, 2020. http://polioeradication.org/wp-content/uploads/2016/07/GPEIStatusReport2014Jan-Jun.pdf

27.

World Health Organization. Statement on the 1st Meeting of the IHR Emergency Committee on the 2014 Ebola outbreak in West Africa. WHO statement. August 8, 2014. Accessed June 10, 2020. https://www.who.int/mediacentre/news/statements/2014/ebola-20140808/en/

28.

Ryan

Carlson

Stewart-Ibarra

, et al. Outbreak of Zika virus infections, Dominica, 2016. Emerg Infect Dis. 2017;23(11):1926–1927. https://doi.org/10.3201/eid2311.171140

29.

Goldberg

. The World Health Organization just declared an Ebola “emergency” in the Democratic Republic of Congo. Here’s what that means. UN Dispatch. July 17, 2019. Accessed June 10, 2020. https://www.undispatch.com/the-world-health-organization-just-declared-an-ebola-emergency-in-the-democratic-republic-of-congo-heres-what-that-means/

30.

WHO Coronavirus Disease (COVID-19) Dashboard. Accessed August 2, 2020. https://covid19.who.int/

31.

Ali

Baloch

Ahmed

, et al. The outbreak of coronavirus disease 2019 (COVID- 19)—an emerging global health threat. J Infect Public Health. 2020;13(4):644–646. https://doi.org/10.1016/j.jiph.2020.02.033

32.

Sullivan

. 3 Charts show where artificial intelligence is making an impact in healthcare right now. Healthcare IT News . December 21, 2018. Accessed November 24, 2020. https://www.healthcareitnews.com/

33.

Davenport

Kalakota

. The potential for artificial intelligence in healthcare. Future Healthc J. 2019;6(2):94–98. https://doi.org/10.7861/futurehosp.6-2-94

34.

Christaki

. New technologies in predicting, preventing and controlling emerging infectious diseases. Virulence. 2015;6(6):558–565. https://doi.org/10.1080/21505594.2015.1040975

35.

Lee

Celik

Logsdon

, et al. A machine learning approach to integrate big data for precision medicine in acute myeloid leukemia. Nat Commun. 2018;9(1):1–3. https://doi.org/10.1038/s41467-017-02465-5

36.

Fakoor

Ladhak

Nazi

, et al. Using deep learning to enhance cancer diagnosis and classification. In Proceedings of the international conference on machine learning 2013 Jun (Vol. 28). New York, USA: ACM.

37.

Vial

Stirling

Field

, et al. The role of deep learning and radiomic feature extraction in cancer-specific predictive modelling: a review. Transl Cancer Res. 2018;7(30):803–816. http://tcr.amegroups.com/article/view/21823

38.

Murff

FitzHenry

Matheny

, et al. Automated identification of postoperative complications within an electronic medical record using natural language processing. JAMA. 2011;306(8):848–855. https://doi.org/10.1001/jama.2011.1204

39.

Darcy

Louie

Roberts

. Machine learning and the profession of medicine. JAMA. 2016;315(6):551–552. https://doi.org/10.1001/jama.2015.18421

40.

Signorini

Segre

Polgreen

. The use of Twitter to track levels of disease activity and public concern in the U.S. during the influenza A H1N1 pandemic. PLoS One. 2011;6(5):e19467. https://doi.org/10.1371/journal.pone.0019467

41.

Eng

Tong

Tan

. Predicting host tropism of Influenza A virus proteins using random forest. BMC Med Genet. 2014;7(Suppl 3):S1. https://doi.org/10.1186/1755-8794-7-S3-S1

42.

Allam

Dey

Jones

. Artificial intelligence (AI) provided early detection of the coronavirus (COVID-19) in China and will influence future urban health policy internationally. AI. 2020;1(2):156–165. https://doi.org/10.3390/ai1020009

43.

44.

Alessa

Faezipour

. A review of influenza detection and prediction through social networking sites. Theor Biol Med Model. 2018;15(1):1–27. https://doi.org/10.1186/s12976-017-0074-5

45.

Chen

Zhang

, et al. Deep learning-based model for detecting 2019 novel coronavirus pneumonia on 2 high-resolution computed tomography: a prospective study. Preprint in medRxiv; March 1, 2020. https://doi.org/10.1101/2020.02.25.20021568

46.

Rawat

Wang

. Deep convolutional neural networks for image classification: a comprehensive review. Neural Comput. 2017;29(9):2352–2449. https://doi.org/10.1162/NECOa00990

47.

Qin

, et al. Artificial intelligence distinguishes COVID-19 from community acquired pneumonia on chest CT. Radiology. 2020. https://doi.org/10.1148/radiol.2020200905

48.

Wang

Wong

. COVID-net: a tailored deep convolutional neural network design for detection of COVID-19 cases from chest radiography images. E-print; March 22, 2020. https://arxiv.org/abs/2003.09871

49.

Wang

Kang

, et al. A deep learning algorithm using CT images to screen for corona virus disease (COVID-19). medRxiv Preprint article; 2020. https://doi.org/10.1101/2020.02.14.20023028

50.

Butt

Gill

Chun

Babu

. Deep learning system to screen coronavirus disease 2019 pneumonia. Appl Intell. 2020:1–7. https://doi.org/10.1007/s10489-020-01714-3

51.

XPHealth. VR-powered telehealth: guided, at-home drug-free recovery. VR/AR Telehealth Process; 2020. Accessed June 10, 2020. https://www.xr.health/

52.

Dias

Torkamani

. Artificial intelligence in clinical and genomic diagnostics. Genome Med. 2019;11(1):1–2. https://doi.org/10.1186/s13073-019-0689-8

53.

Alimadadi

Aryal

Manandhar

, et al. Artificial intelligence and machine learning to fight COVID-19. Physiol Genomics. 2020;52(4):200–202. https://doi.org/10.1152/physiolgenomics.00029.2020

54.

Zhang

Saravanan

Yang

, et al. Deep learning based drug screening for novel coronavirus 2019-nCov. Interdiscip Sci Comput Life Sci. 2020;12:368–376. https://doi.org/10.1007/s12539-020-00376-6

55.

Park

Casey

Joshi

, et al. Emergence of new disease: how can artificial intelligence help? Trends Mol Med. 2020;26(7):627–629. https://doi.org/10.1016/j.molmed.2020.04.007

56.

Akhtar

Kraemer

MUG

Gardner

. A dynamic neural network model for predicting risk of Zika in real time. BMC Med. 2019;17(1):1–6. https://doi.org/10.1186/s12916-019-1389-3

57.

McCall

. COVID-19 and artificial intelligence: protecting health-care workers and curbing the spread. Lancet Digit Health. 2020;2(4):e166–e167. https://doi.org/10.1016/S2589-7500(20)30054-6

58.

Wang

Kang

, et al. Regional influenza prediction with sampling twitter data and PDE model. Int J Environ Res Public Health. 2020;17(3):678. https://doi.org/10.3390/ijerph17030678

59.

Guo

Wang

, et al. Host and infectivity prediction of Wuhan 2019 novel coronavirus using deep learning algorithm. bioRxiv Preprint; January 24, 2020. https://doi.org/10.1101/2020.01.21.914044

60.

Yan

Zhang

Goncalves

, et al. A machine learning-based model for survival prediction in patients with severe COVID-19 infection. medRxiv Preprint; March 17, 2020. https://doi.org/10.1101/2020.02.27.20028027

61.

Lazer

Kennedy

King

Vespignani

. The parable of Google flu: traps in big data analysis. Science. 2014;343(6176):1203–1205. https://doi.org/10.1126/science.1248506

62.

Dugas

Hsieh

Levin

, et al. Google flu trends: correlation with emergency department influenza rates and crowding metrics. Clin Infect Dis. 2012;54(4):463–469. https://doi.org/10.1093/cid/cir883

63.

Fong

Dey

, et al. Finding an accurate early forecasting model from small dataset: a case of 2019-nCoV novel coronavirus outbreak. Int J Interact Multimed Artif Intell. 2020;6(1):132–140. https://doi.org/10.9781/ijimai.2020.02.002

64.

Amazon Web Services (AWS). CORD-19 Search. AWS Search Engine. Accessed June 10, 2020. https://cord19.aws/

65.

C3.ai. AI Software Company, Redwood City, CA. Accessed June 10, 2020. https://c3.ai/

66.

Broad

. A.I. versus the coronavirus. NY Times News . March 26, 2020. Accessed June 10, 2020. https://www.nytimes.com/2020/03/26/science/ai-versus-the-coronavirus.html

67.

Beck

Shin

Choi

, et al. Predicting commercially available antiviral drugs that may act on the novel coronavirus (SARS-CoV-2) through a drug- target interaction deep learning model. Comput Struct Biotechnol J. 2020;18:784–790. https://doi.org/10.1016/j.csbj.2020.03.025

68.

Benevolent AI. AI company, London, UK. Accessed June 10, 2020. https://www.benevolent.com/covid-19

69.

Fleming

. Computer-calculated compounds: researchers are deploying artificial intelligence to discover drugs. Nature. 2018;557:S55–S57.

70.

Segler

Preuss

Waller

. Planning chemical syntheses with deep neural networks and symbolic AI. Nature. 2018;555(7698):604–610. https://doi.org/10.1038/nature25978

71.

Laksham

. Unmanned aerial vehicle (drones) in public health: a SWOT analysis. J Family Med Prim Care. 2019;8(2):342–346. https://doi.org/10.4103/jfmpc.jfmpc41318

72.

Enabling rapid and open access to epidemic and pandemic virus data. Accessed June 5, 2020. https://www.gisaid.org/about-us/mission/

73.

Zhao

, et al. A new coronavirus associated with human respiratory disease in China. Nature. 2020;579(7798):265–269. https://doi.org/10.1038/s41586-020-2008-3

74.

Coronavirus Genome Sequence: 2019-nCoV/SARS-CoV-2. Accessed June 5, 2020. https://www.kaggle.com/paultimothymooney/coronavirus-genome-sequence

75.

COVID-19 Open Research Dataset Challenge (CORD-19). Accessed June 5, 2020. https://www.kaggle.com/allen-institute-for-ai/CORD-19-research-challenge

76.

Haslett

. COVID-19-growth. Accessed June 5, 2020. https://github.com/willhaslett/covid-19-growth

77.

Lee

. Data Science for COVID-19. Accessed June 5, 2020. https://github.com/ThisIsIsaac/Data-Science-for-COVID-19

78.

Cohen

Morrison

Dao

, et al. COVID-19 image data collection: prospective predictions are the future. arXiv:2006.11988; 2020. Accessed June 5, 2020. https://github.com/ieee8023/covid-chestxray-dataset

79.

Coronavirus (COVID19) Tweets Dataset. Kaggle Repository. Accessed June 5, 2020. https://www.kaggle.com/smid80/coronavirus-covid19-tweets

80.

COVID-19 Coronavirus data. EU Open Data Portal. Accessed July 5, 2020. http://data.europa.eu/88u/dataset/covid-19-coronavirus-data. https://doi.org/10.2906/101099100099/1

81.

COVID-19 Pandemic. Accessed July 5, 2020. https://data.humdata.org/event/covid-19

82.

Novel Coronavirus (COVID-19) Cases Data. Accessed July 5, 2020. https://data.humdata.org/dataset/novel-coronavirus-2019-ncov-cases

83.

COVID-19 Data Repository. Accessed July 5, 2020. https://github.com/CSSEGISandData/COVID-19

84.

Data on COVID-19 (Coronavirus). Accessed June 5, 2020. https://github.com/owid/covid-19-data/tree/master/public/data

85.

Coronavirus (COVID-19) Tweets Dataset. Accessed June 5, 2020. https://ieee-dataport.org/open-access/coronavirus-covid-19-tweets-dataset

86.

COVID-19 Coronavirus Pandemic. Accessed July 5, 2020. https://www.worldometers.info/coronavirus/

87.

Mooney

. Chest X-ray Images (Pneumonia). Kaggle Repository; 2018. Accessed July 4, 2020. https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia

88.

ImageNet. Stanford Vision Lab, Stanford University, Princeton University; 2016. Accessed July 5, 2020. http://image-net.org/

89.

Russakovsky

Deng

, et al. ImageNet large scale visual recognition challenge. Int J Comput Vis. 2015;115(3):211–252. https://doi.org/10.1007/s11263-015-0816-y

90.

HIV sequence database. Accessed July 5, 2020. http://hiv.lanl.gov/catnap

91.

Yoon

Macke

West

, et al. CATNAP: a tool to compile, analyze and tally neutralizing antibody panels. Nucleic Acids Res. 2015;43(W1):W213–W219. https://doi.org/10.1093/nar/gkv404

92.

Berman

Westbrook

Feng

, et al. The protein data bank. Nucleic Acids Res. 2000;28(1):235–242. https://doi.org/10.1093/nar/28.1.235

93.

COVID-19 Data Lake Sample Dashboard. Accessed July 5, 2020. https://covid19-lake.s3.us-east-2.amazonaws.com/dashboard.html?dashboardName=COVID-19

94.

James

Witten

Hastie

, et al. An Introduction to Statistical Learning with Applications in R. 1st ed. Springer; 2013. ISBN 978-1-4614-7138-7

95.

Goodfellow

Bengio

Courville

. Deep Learning. 1st ed. The MIT Press; 2016. ISBN: 9780262035613

96.

Manning

Schutze

. Foundations of Statistical Natural Language Processing. 2nd ed . The MIT Press; 2000. ISBN 0-262-13360-l

97.

Hesamian

Jia

, et al. Deep learning techniques for medical image segmentation: achievements and challenges. J Digit Imaging. 2019;32(4):582–596. https://doi.org/10.1007/s10278-019-00227-x

98.

Akagi

Nakamura

Higaki

, et al. Correction to: deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT. Eur Radiol. 2019;29(11):4526–4527. https://doi.org/10.1007/s00330-019-06249-x

99.

Nardelli

Jimenez-Carretero

Bermejo-Pelaez

, et al. Pulmonary artery–vein classification in CT images using deep learning. IEEE Trans Med Imaging. 2018;37(11):2428–2440. https://doi.org/10.1109/TMI.2018.2833385

100.

Zhu

Huang

Zeng

, et al. Anatomynet: deep learning for fast and fully automated whole-volume segmentation of head and neck anatomy. Med Phys. 2019;46(2):576–589. https://doi.org/10.1002/mp.13300

101.

Saba

Sameh

El-Affendi

, et al. Brain tumor detection using fusion of hand crafted and deep learning features. Cogn Syst Res. 2020;59:221–230. https://doi.org/10.1016/j.cogsys.2019.09.007

102.

Dorj

Lee

Choi

, et al. The skin cancer classification using deep convolutional neural network. Multimed Tools Appl. 2018;77(8):9909–9924. https://doi.org/10.1007/s11042-018-5714-1

103.

Kassani

. A comparative study of deep learning architectures on melanoma detection. Tissue Cell. 2019;58:76–83. https://doi.org/10.1016/j.tice.2019.04.009

104.

Ribli

Horváth

Unger

, et al. Detecting and classifying lesions in mammograms with deep learning. Sci Rep. 2018;8(1):1–7. https://doi.org/10.1038/s41598-018-22437-z

105.

Celik

Talo

Yildirim

, et al. Automated invasive ductal carcinoma detection based using deep transfer learning with whole-slide images. Pattern Recognit Lett. 2020;133:232–239. https://doi.org/10.1016/j.patrec.2020.03.011

106.

Pham

Nguyen

Huynh-The

Hwang

Pathirana

. Artificial intelligence (AI) and big data for coronavirus (COVID-19) pandemic: a survey on the state-of-the-arts. IEEE Access. 2020;8:130820–130839. https://doi.org/10.1109/ACCESS.2020.3009328

107.

Maghdid

Ghafoor

Sadiq

, et al. A novel AI-enabled framework to diagnose coronavirus COVID-19 using smartphone embedded sensors: design study. arXiv Preprint; 2020. https://arxiv.org/abs/2003.07434

108.

Rao

ASS

Vazquez

. Identification of COVID-19 can be quicker through artificial intelligence framework using a mobile phone-based survey in the populations when cities/towns are under quarantine. Infect Control Hosp Epidemiol. 2020;41(7):826–830. https://doi.org/10.1017/ice.2020.61

109.

Barstugan

Ozkaya

Ozturk

. Coronavirus (COVID-19) classification using CT images by machine learning methods. arXiv Preprint; 2020. https://arxiv.org/abs/2003.09424

110.

Abbas

Abdelsamea

Gaber

. Classification of COVID-19 in chest X-ray images using DeTraC deep convolutional neural network. arXiv Preprint; 2020. https://arxiv.org/abs/2003.13815

111.

Apostolopoulos

Mpesiana

. Covid-19: automatic detection from X-ray images utilizing transfer learning with convolutional neural networks. Phys Eng Sci Med. 2020;43(2):635–640. https://doi.org/10.1007/s13246-020-00865-4

112.

Narin

Kaya

Pamuk

. Automatic detection of coronavirus disease (COVID-19) using X-ray images and deep convolutional neural networks. arXiv Preprint; 2020. https://arxiv.org/abs/2003.10849

113.

Ghoshal

Tucker

. Estimating uncertainty and interpretability in deep learning for coronavirus (COVID-19) detection. arXiv Preprint 2020. https://arxiv.org/abs/2003.10769

114.

Dai

Zhang

, et al. CT imaging and differential diagnosis of COVID-19. Can Assoc Radiol J. 2020;71(2):195–200. https://doi.org/10.1177/0846537120913033

115.

Wang

Wong

. COVID-net: a tailored deep convolutional neural network design for detection of COVID-19 cases from chest x-ray images. arXiv Preprint; 2020. https://arxiv.org/abs/2003.09871

116.

Shen

Van DenHengel

. Wider or deeper: revisiting the ResNet model for visual recognition. Pattern Recognit. 2019;90:119–133. https://doi.org/10.1016/j.patcog.2019.01.006

117.

Ahn

Kim

Ahn

, et al. Correction: a deep learning model for the detection of both advanced and early glaucoma using fundus photography. PLoS One. 2019;14(1):e0211579. https://doi.org/10.1371/journal.pone.0211579

118.

Chollet

. Xception: deep learning with depth wise separable convolutions. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI; 2017:1800–1807. https://doi.org/10.1109/CVPR.2017.195

119.

Simonyan

Zisserman

. Very deep convolutional networks for large-scale image recognition. arXiv Preprint; 2015. https://arxiv.org/abs/1409.1556

120.

Horry

Chakraborty

Paul

, et al. X-ray image based COVID-19 detection using pre-trained deep learning models. engrXiv Preprint; 2020. https://doi.org/10.31224/osf.io/wx89s

121.

Byra

Styczynski

Szmigielski

, et al. Transfer learning with deep convolutional neural network for liver steatosis assessment in ultrasound images. Int J Comput Assist Radiol Surg. 2018;13(12):1895–1903. https://doi.org/10.1007/s11548-018-1843-2

122.

Haque

Gandy

Abdelgawad

. Automatic detection of COVID-19 from chest X-ray images with convolutional neural networks. In: 2020 International Conference on Computing, Electronics & Communications Engineering (ICCECE), August 17, 2020:125–130. IEEE.

123.

Song

Zheng

, et al. Deep learning enables accurate diagnosis of novel coronavirus (COVID-19) with CT images. medRxiv; January 1, 2020.

124.

Jiang

, et al. A deep learning system to screen novel coronavirus disease 2019 pneumonia. Engineering. 2020;6(10):1122–1129.

125.

Ozturk

Talo

Yildirim

Baloglu

Yildirim

Acharya

. Automated detection of COVID-19 cases using deep neural networks with X-ray images. Comput Biol Med. 2020;121:103792.

126.

Jiang

Zhi

, et al. Artificial intelligence in healthcare: past, present and future. Stroke Vasc Neurol. 2017;2(4):230–243. https://doi.org/10.1136/svn-2017-000101

127.

Rahmatizadeh

Valizadeh-Haghi

Dabbagh

. The role of artificial intelligence in management of critical COVID-19 patients. J Cell Mol Anesthesia. 2020;5(1):16–22. https://doi.org/10.22037/jcma.v5i1.29752

128.

Davoudi

Malhotra

Shickel

, et al. Intelligent ICU for autonomous patient monitoring using pervasive sensing and deep learning. Sci Rep. 2019;9(1):1–3. https://doi.org/10.1038/s41598-019-44004-w

129.

Gutierrez

. Artificial intelligence in the intensive care unit. Crit Care. 2020;24:1–9. https://doi.org/10.1186/s13054-020-2785-y

130.

Vaishya

Javid

Khan

, et al. Artificial intelligence (AI) applications for COVID-19 pandemic. Diabetes Metab Syndr. 2020;14(4):337–339. https://doi.org/10.1016/j.dsx.2020.04.012

131.

Singh

Mishra

. Application of box- Jenkins method and artificial neural network procedure for time series forecasting of prices. Stat Transition New Ser. 2015;16(1):83–96.

132.

Arun

Krishna

Arunkumar

, et al. Exploratory boosted feature selection and neural network framework for depression classification. Int J Interact Multimed Artif Intell. 2018;5(3):61–71. https://doi.org/10.9781/ijimai.2018.10.001

133.

Magar

Yadav

Farimani

. Potential neutralizing antibodies discovered for novel corona virus using machine learning. bioRxiv Preprint; 2020. https://doi.org/10.1101/2020.03.14.992156

134.

GitHub and SourceForge. RDKit: open-source cheminformatics software. Accessed July 5, 2020. https://www.rdkit.org/

135.

Lai

C-C

Shih

T-P

W-C

, et al. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV- 2) and coronavirus disease-2019 (COVID-19): the epidemic and the challenges. Int J Antimicrob Agents. 2020;55(3):105924. https://doi.org/10.1016/j.ijantimicag.2020.105924

136.

Mamoshina

Vieira

Putin

Zhavoronkov

. Applications of deep learning in biomedicine. Mol Pharm. 2016;13(5):1445–1454. https://doi.org/10.1021/acs.molpharmaceut.5b00982

137.

Amazon Kendra. Amazon Web Services, Inc.; 2020. Accessed on July 5, 2020. https://aws.amazon.com/kendra/

138.

Kass-Hout

Snively

. AWS Launches machine learning enabled search capabilities for COVID-19 dataset. AWS Public Sector Blog ; April 20, 2020. Accessed June 10, 2020. https://aws.amazon.com/blogs/publicsector/aws-launches-machine-learning-enabled-search-capabilities-covid-19-dataset/

139.

Yang

Liu

Chen

Tong

. Federated machine learning: concept and applications. ACM Trans Intell Syst Technol. 2019;10(2):1–9.

140.

Gao

Liu

Wang

, et al. A survey of incentive mechanisms for participatory sensing. IEEE Commun Surv Tutorials. 2015;17(2):918–943.

141.

Rees

. AI And cloud computing used to develop COVID-19 vaccine. Drug Target Review ; April 8, 2020. https://www.drugtargetreview.com/news/59650/ai-and-cloud-computing-used-to-develop-covid-19-vaccine/

142.

Chamola

Hassija

Gupta

Guizani

. A comprehensive review of the COVID-19 pandemic and the role of IoT, drones, AI, blockchain, and 5G in managing its impact. IEEE Access. 2020;8:90225–90265. https://doi.org/10.1109/ACCESS.2020.2992341

A Comprehensive Study of Artificial Intelligence and Machine Learning Approaches in Confronting the Coronavirus (COVID-19) Pandemic

Abstract

Keywords

Background

Machine Learning and Deep Learning

Natural Language Processing

Expert Systems

Robots and Robotic Process

Understanding and Detecting

Diagnosis, Genome Analysis, and Treatment

Prediction and Forecasting the Outbreak

Speeding up Research and Drug Discovery

Public Health Management and Services

Related Datasets

Artificial Intelligence Approaches in Health Care

Machine Learning Models for Detection and Diagnosis

Machine Learning and Natural Language Processing Model for Clinical Decision Making and Management

Artificial Intelligence System for Identifying COVID-19 Symptoms

Machine Learning-based Forecasting Methods

Machine Learning for Accelerating Drug Development

Machine Learning-based Search Engine

Discussion and Future Directions

Conclusion

Footnotes

Acknowledgments

Declaration of Conflicting Interests

Funding

ORCID iD

Supplemental Material

Author Biographies

References